Show HN: Starling – Wearable to measure number of words spoken to your child | Hacker News Reader

Show HN: Starling – Wearable to measure number of words spoken to your child

versame.com

16 pointsmbesto11 comments

I'm also concerned about the privacy implications, but think it would be neat to hack this thing into a personal Robot9000 machine - though I don't know how it would punish you.

http://blog.xkcd.com/2008/01/14/robot9000-and-xkcd-signal-at...

From their website:

  "...Starling doesn’t record anything you say – it simply counts words as they pass by."

For privacy concerned among us it would be nice to see them elaborate a bit more on privacy. That being said, this device looks amazing.

It is quite amazing how the founders were able to translate serious Stanford research into a simple and affordable product.

Hi I'm Chris, one of the creators of the Starling. Here's a link to the Stanford research on early childhood language processing if you or anyone else is interested:

https://web.stanford.edu/group/langlearninglab/cgi-bin/publi...

There's more from other research institutions, but Anne Fernald from Stanford's Language Learning Lab is has been an advisor of ours since the beginning.

On affordability, it's not lost on us that the Starling costs less than a single day's tuition at Stanford. We hope to make it more affordable still.

Is the analytics on the phone or would the user be trusting you with a recording of every private moment of their lives?

The graphic says that the microphone picks up words without recording: http://www.versame.com/wp-content/uploads/2015/10/exploded-s...

So I'm guessing they have some algorithm that can identify breaks between words and count based on that.

Awesome. That'll teach me for just being cynical after just skimming over the front page of the mobile site.

That said, I'm still pretty skeptical about the grandiose claims about the number of words a child hearing per day and IQ. I strongly suspect that hearing large numbers of words before the age of 4 is not causing increases in intelligence. My guess is that it's just a strong correlate to a bunch of harder to measure factors.

Still, the fact that I wouldn't be buying this at the cost of my privacy makes me way more likely to get this thing. Whenever I'm learning a language, I do wonder about how long it takes to hear most of the words you need to get along in daily life, and how often you hear them. Even if it doesn't give you any kind of list (shamefully I still haven't looked at the details), it'd be interesting to just know the raw number.

Hi there, I'm one of the creators of the Starling and I can reassure you that privacy is as sacred to us as it is to you - nothing is recorded. The entire focus is on volume of words. That's not to say quality (diversity, length, conversation) isn't important but research has found them to be very highly correlated. Correlation or causation, encouraging more language and engagement between parents and children in a world of distracting electronic devices can't be a bad thing? That's what our Beta testers have told us (you can listen to the video of testimonials on the site), that they are more engaged and focused on their children and their families are happier as a result. Can't complain about that!

It should be able to distinguish words. I bet that on some days, the half of the words my 10 month old hears are "No!".

Hi I'm one of Starling's creators. We decided to err on the side of privacy. The good news is that research has shown that quality and quantity are really highly correlated - if you choose to talk more to your child's you're highly likely to say positive things.

That's a really interesting fact. Thanks for sharing!

Reply on news.ycombinator.com